+2009-09-25 Revital Eres <ERES@il.ibm.com>
+
+ * gcc.target/powerpc/vsx-vectorize-8.c: New test.
+ * gcc.target/powerpc/vsx-vectorize-1.c: Likewise.
+ * gcc.target/powerpc/vsx-vectorize-2.c: Likewise.
+ * gcc.target/powerpc/vsx-vectorize-3.c: Likewise.
+ * gcc.target/powerpc/vsx-vectorize-4.c: Likewise.
+ * gcc.target/powerpc/vsx-vectorize-5.c: Likewise.
+ * gcc.target/powerpc/vsx-vectorize-6.c: Likewise.
+ * gcc.target/powerpc/vsx-vectorize-7.c: Likewise.
+ * gcc.dg/vect/vect.exp: Add Power7 flags.
+ * g++.dg/vect/vect.exp: Likewise.
+ * gfortran.dg/vect/vect.exp: Likewise.
+
2009-09-24 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/41459
}
lappend DEFAULT_VECTCFLAGS "-maltivec"
+ if [check_vsx_hw_available] {
+ lappend DEFAULT_VECTCFLAGS "-mvsx" "-mno-allow-movmisalign"
+ }
+
if [check_vmx_hw_available] {
set dg-do-what-default run
} else {
}
lappend DEFAULT_VECTCFLAGS "-maltivec"
+ if [check_vsx_hw_available] {
+ lappend DEFAULT_VECTCFLAGS "-mvsx" "-mno-allow-movmisalign"
+ }
+
if [check_vmx_hw_available] {
set dg-do-what-default run
} else {
--- /dev/null
+/* { dg-do compile { target { powerpc*-*-* && lp64 } } } */
+/* { dg-skip-if "" { powerpc*-*-darwin* } { "*" } { "" } } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-mcpu=power7 -O2 -ftree-vectorize -fno-vect-cost-model -fdump-tree-vect-details" } */
+
+/* Taken from vect/vect-align-1.c. */
+#include <stdlib.h>
+#include <stdarg.h>
+
+/* Compile time known misalignment. Cannot use loop peeling to align
+ the store. */
+
+#define N 16
+
+struct foo {
+ char x;
+ int y[N];
+} __attribute__((packed));
+
+int x[N] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15};
+
+__attribute__ ((noinline)) int
+main1 (struct foo * __restrict__ p)
+{
+ int i;
+
+ for (i = 0; i < N; i++)
+ {
+ p->y[i] = x[i];
+ }
+
+ /* check results: */
+ for (i = 0; i < N; i++)
+ {
+ if (p->y[i] != x[i])
+ abort ();
+ }
+ return 0;
+}
+
+
+int main (void)
+{
+ int i;
+ struct foo *p = malloc (2*sizeof (struct foo));
+
+ main1 (p);
+ return 0;
+}
+
+/* { dg-final { scan-tree-dump-times "Vectorizing an unaligned access" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using versioning" 1 "vect" } } */
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" } } */
+/* { dg-final { cleanup-tree-dump "vect" } } */
--- /dev/null
+/* { dg-do compile { target { powerpc*-*-* && lp64 } } } */
+/* { dg-skip-if "" { powerpc*-*-darwin* } { "*" } { "" } } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-mcpu=power7 -O2 -ftree-vectorize -fno-vect-cost-model -fdump-tree-vect-details" } */
+
+/* Taken from vect/vect-95.c. */
+#include <stdlib.h>
+#include <stdarg.h>
+
+#define N 256
+
+__attribute__ ((noinline))
+void bar (float *pd, float *pa, float *pb, float *pc)
+{
+ int i;
+
+ /* check results: */
+ for (i = 0; i < N; i++)
+ {
+ if (pa[i] != (pb[i] * pc[i]))
+ abort ();
+ if (pd[i] != 5.0)
+ abort ();
+ }
+
+ return;
+}
+
+
+__attribute__ ((noinline)) int
+main1 (int n, float * __restrict__ pd, float * __restrict__ pa, float * __restrict__ pb, float * __restrict__ pc)
+{
+ int i;
+
+ for (i = 0; i < n; i++)
+ {
+ pa[i] = pb[i] * pc[i];
+ pd[i] = 5.0;
+ }
+
+ bar (pd,pa,pb,pc);
+
+ return 0;
+}
+
+int main (void)
+{
+ int i;
+ float a[N] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
+ float d[N+1] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
+ float b[N] = {0,3,6,9,12,15,18,21,24,27,30,33,36,39,42,45,48,51,54,57};
+ float c[N] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19};
+
+ main1 (N,&d[1],a,b,c);
+ main1 (N-2,&d[1],a,b,c);
+
+ return 0;
+}
+
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using peeling" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using versioning" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Vectorizing an unaligned access" 4 "vect" } } */
+/* { dg-final { cleanup-tree-dump "vect" } } */
--- /dev/null
+/* { dg-do compile { target { powerpc*-*-* && lp64 } } } */
+/* { dg-skip-if "" { powerpc*-*-darwin* } { "*" } { "" } } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-mcpu=power7 -O2 -ftree-vectorize -fno-vect-cost-model -fdump-tree-vect-details" } */
+
+/* Taken from vect/vect-95.c. */
+#include <stdlib.h>
+#include <stdarg.h>
+
+#define N 256
+
+__attribute__ ((noinline))
+void bar (short *pa, short *pb, short *pc)
+{
+ int i;
+
+ /* check results: */
+ for (i = 0; i < N; i++)
+ {
+ if (pa[i] != (pb[i] * pc[i]))
+ abort ();
+ }
+
+ return;
+}
+
+
+__attribute__ ((noinline)) int
+main1 (int n, short * __restrict__ pa, short * __restrict__ pb, short * __restrict__ pc)
+{
+ int i;
+
+ for (i = 0; i < n; i++)
+ {
+ pa[i] = pb[i] * pc[i];
+ }
+
+ bar (pa,pb,pc);
+
+ return 0;
+}
+
+int main (void)
+{
+ int i;
+ short a[N] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
+ short b[N] = {0,3,6,9,12,15,18,21,24,27,30,33,36,39,42,45,48,51,54,57};
+ short c[N] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19};
+
+ main1 (N,a,b,c);
+ main1 (N-2,a,b,c);
+
+ return 0;
+}
+
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using peeling" 1 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using versioning" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Vectorizing an unaligned access" 2 "vect" } } */
+/* { dg-final { cleanup-tree-dump "vect" } } */
--- /dev/null
+/* { dg-do compile { target { powerpc*-*-* && lp64 } } } */
+/* { dg-skip-if "" { powerpc*-*-darwin* } { "*" } { "" } } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-mcpu=power7 -O2 -ftree-vectorize -fno-vect-cost-model -fdump-tree-vect-details" } */
+
+/* Taken from vect/vect-95.c. */
+#include <stdlib.h>
+#include <stdarg.h>
+
+#define N 256
+
+__attribute__ ((noinline))
+void bar (double *pa, double *pb, double *pc)
+{
+ int i;
+
+ /* check results: */
+ for (i = 0; i < N; i++)
+ {
+ if (pa[i] != (pb[i] * pc[i]))
+ abort ();
+ }
+
+ return;
+}
+
+
+__attribute__ ((noinline)) int
+main1 (int n, double * __restrict__ pa, double * __restrict__ pb, double * __restrict__ pc)
+{
+ int i;
+
+ for (i = 0; i < n; i++)
+ {
+ pa[i] = pb[i] * pc[i];
+ }
+
+ bar (pa,pb,pc);
+
+ return 0;
+}
+
+int main (void)
+{
+ int i;
+ double a[N] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
+ double b[N] = {0,3,6,9,12,15,18,21,24,27,30,33,36,39,42,45,48,51,54,57};
+ double c[N] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19};
+
+ main1 (N,a,b,c);
+ main1 (N-2,a,b,c);
+
+ return 0;
+}
+
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using peeling" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using versioning" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Vectorizing an unaligned access" 3 "vect" } } */
+/* { dg-final { cleanup-tree-dump "vect" } } */
--- /dev/null
+/* { dg-do compile { target { powerpc*-*-* && lp64 } } } */
+/* { dg-skip-if "" { powerpc*-*-darwin* } { "*" } { "" } } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-mcpu=power7 -O2 -ftree-vectorize -fno-vect-cost-model -fdump-tree-vect-details" } */
+
+/* Taken from vect/vect-95.c. */
+#include <stdlib.h>
+#include <stdarg.h>
+
+#define N 256
+
+__attribute__ ((noinline))
+void bar (char *pa, char *pb, char *pc)
+{
+ int i;
+
+ /* check results: */
+ for (i = 0; i < N; i++)
+ {
+ if (pa[i] != (pb[i] + pc[i]))
+ abort ();
+ }
+
+ return;
+}
+
+
+__attribute__ ((noinline)) int
+main1 (int n, char * __restrict__ pa, char * __restrict__ pb, char * __restrict__ pc)
+{
+ int i;
+
+ for (i = 0; i < n; i++)
+ {
+ pa[i] = pb[i] + pc[i];
+ }
+
+ bar (pa,pb,pc);
+
+ return 0;
+}
+
+int main (void)
+{
+ int i;
+ char a[N] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
+ char b[N] = {0,3,6,9,12,15,18,21,24,27,30,33,36,39,42,45,48,51,54,57};
+ char c[N] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19};
+
+ main1 (N,a,b,c);
+ main1 (N-2,a,b,c);
+
+ return 0;
+}
+
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using peeling" 1 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using versioning" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Vectorizing an unaligned access" 2 "vect" } } */
+/* { dg-final { cleanup-tree-dump "vect" } } */
--- /dev/null
+/* { dg-do compile { target { powerpc*-*-* && lp64 } } } */
+/* { dg-skip-if "" { powerpc*-*-darwin* } { "*" } { "" } } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-mcpu=power7 -O2 -ftree-vectorize -fno-vect-cost-model -fdump-tree-vect-details" } */
+
+/* Taken from vect/vect-95.c. */
+#include <stdlib.h>
+#include <stdarg.h>
+
+#define N 256
+
+__attribute__ ((noinline))
+void bar (double *pd, double *pa, double *pb, double *pc)
+{
+ int i;
+
+ /* check results: */
+ for (i = 0; i < N; i++)
+ {
+ if (pa[i] != (pb[i] * pc[i]))
+ abort ();
+ if (pd[i] != 5.0)
+ abort ();
+ }
+
+ return;
+}
+
+
+__attribute__ ((noinline)) int
+main1 (int n, double * __restrict__ pd, double * __restrict__ pa, double * __restrict__ pb, double * __restrict__ pc)
+{
+ int i;
+
+ for (i = 0; i < n; i++)
+ {
+ pa[i] = pb[i] * pc[i];
+ pd[i] = 5.0;
+ }
+
+ bar (pd,pa,pb,pc);
+
+ return 0;
+}
+
+int main (void)
+{
+ int i;
+ double a[N] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
+ double d[N+1] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
+ double b[N] = {0,3,6,9,12,15,18,21,24,27,30,33,36,39,42,45,48,51,54,57};
+ double c[N] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19};
+
+ main1 (N,&d[1],a,b,c);
+ main1 (N-2,&d[1],a,b,c);
+
+ return 0;
+}
+
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using peeling" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using versioning" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Vectorizing an unaligned access" 4 "vect" } } */
+/* { dg-final { cleanup-tree-dump "vect" } } */
--- /dev/null
+/* { dg-do compile { target { powerpc*-*-* && lp64 } } } */
+/* { dg-skip-if "" { powerpc*-*-darwin* } { "*" } { "" } } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-mcpu=power7 -O2 -ftree-vectorize -fno-vect-cost-model -fdump-tree-vect-details" } */
+
+/* Taken from vect/vect-95.c. */
+#include <stdlib.h>
+#include <stdarg.h>
+
+#define N 256
+
+__attribute__ ((noinline))
+void bar (int *pd, int *pa, int *pb, int *pc)
+{
+ int i;
+
+ /* check results: */
+ for (i = 0; i < N; i++)
+ {
+ if (pa[i] != (pb[i] * pc[i]))
+ abort ();
+ if (pd[i] != 5.0)
+ abort ();
+ }
+
+ return;
+}
+
+
+__attribute__ ((noinline)) int
+main1 (int n, int * __restrict__ pd, int * __restrict__ pa, int * __restrict__ pb, int * __restrict__ pc)
+{
+ int i;
+
+ for (i = 0; i < n; i++)
+ {
+ pa[i] = pb[i] * pc[i];
+ pd[i] = 5.0;
+ }
+
+ bar (pd,pa,pb,pc);
+
+ return 0;
+}
+
+int main (void)
+{
+ int i;
+ int a[N] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
+ int d[N+1] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
+ int b[N] = {0,3,6,9,12,15,18,21,24,27,30,33,36,39,42,45,48,51,54,57};
+ int c[N] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19};
+
+ main1 (N,&d[1],a,b,c);
+ main1 (N-2,&d[1],a,b,c);
+
+ return 0;
+}
+
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using peeling" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using versioning" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Vectorizing an unaligned access" 4 "vect" } } */
+/* { dg-final { cleanup-tree-dump "vect" } } */
--- /dev/null
+/* { dg-do compile { target { powerpc*-*-* && lp64 } } } */
+/* { dg-skip-if "" { powerpc*-*-darwin* } { "*" } { "" } } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-mcpu=power7 -O2 -ftree-vectorize -fno-vect-cost-model -fdump-tree-vect-details" } */
+
+/* Taken from vect/vect-95.c. */
+#include <stdlib.h>
+#include <stdarg.h>
+
+#define N 256
+
+__attribute__ ((noinline))
+void bar (short *pd, short *pa, short *pb, short *pc)
+{
+ int i;
+
+ /* check results: */
+ for (i = 0; i < N; i++)
+ {
+ if (pa[i] != (pb[i] * pc[i]))
+ abort ();
+ if (pd[i] != 5.0)
+ abort ();
+ }
+
+ return;
+}
+
+
+__attribute__ ((noinline)) int
+main1 (int n, short * __restrict__ pd, short * __restrict__ pa, short * __restrict__ pb, short * __restrict__ pc)
+{
+ int i;
+
+ for (i = 0; i < n; i++)
+ {
+ pa[i] = pb[i] * pc[i];
+ pd[i] = 5.0;
+ }
+
+ bar (pd,pa,pb,pc);
+
+ return 0;
+}
+
+int main (void)
+{
+ int i;
+ short a[N] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
+ short d[N+1] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
+ short b[N] = {0,3,6,9,12,15,18,21,24,27,30,33,36,39,42,45,48,51,54,57};
+ short c[N] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19};
+
+ main1 (N,&d[1],a,b,c);
+ main1 (N-2,&d[1],a,b,c);
+
+ return 0;
+}
+
+/* { dg-final { scan-tree-dump-times "vectorized 1 loops" 1 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using peeling" 0 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Alignment of access forced using versioning" 2 "vect" } } */
+/* { dg-final { scan-tree-dump-times "Vectorizing an unaligned access" 2 "vect" } } */
+/* { dg-final { cleanup-tree-dump "vect" } } */
}
lappend DEFAULT_VECTCFLAGS "-maltivec"
+ if [check_vsx_hw_available] {
+ lappend DEFAULT_VECTCFLAGS "-mvsx" "-mno-allow-movmisalign"
+ }
+
if [check_vmx_hw_available] {
set dg-do-what-default run
} else {